Global Leadership Project

TL;DR: Key Findings from the Country Rankings

• No country achieves perfect descriptive representation. In our data, every country scores below what we would expect under random selection; representation gap values range from about −0.27 to −0.04, meaning national leaders are consistently less diverse than their citizens.
• Larger political bodies show smaller representation gaps. Legislatures and leadership bodies with more seats are better able to mirror population diversity than very small elites, improving descriptive representation across gender, ethnicity, religion, and language.
• Highly fractionalized societies are harder to represent descriptively. Countries with many distinct ethnic, religious, or language groups tend to have larger representation gaps than more homogeneous societies, even when overall scores look similar on the surface.
• Gender gaps are the most severe. Across most countries, women are more underrepresented in national leadership than major ethnic, religious, or language groups, making the gender representation gap the single largest shortfall in descriptive representation.
• Best use case: benchmarking representation by country and identity dimension. These rankings are most useful for comparing countries, tracking change over time, and evaluating whether institutional reforms reduce representation gaps rather than for grading democracies overall.

Download Country-Level Representation Scores and Raw Data

You can access both the country-level representation scores and the underlying Global Leadership Project data to reproduce and extend the analysis. Use these files to run your own models, create visualizations, or integrate descriptive representation benchmarks into other research projects.

• Download country-level representation scores (CSV)
• Download raw Global Leadership Project data and replication code (GlobalLeadershipProject_v1)
• Read the American Political Science Review article “The Composition of Descriptive Representation”

What Is the Representation Gap Index?

The representation gap index captures how far a country’s political leadership falls short of perfectly mirroring the demographic makeup of its population. For each identity dimension—gender, ethnicity, religion, and language—we compare the observed share of groups in national leadership to what we would expect if officeholders were selected at random from the adult population. The closer a country comes to that random-selection benchmark, the smaller its representation gap.

The dataset reports both dimension-specific representation indices and summary measures for each country:

• Overall: a composite descriptive representation score that summarizes how closely leaders match citizens across all four identity dimensions.
• Ethnicity, Gender, Religion, Language: representation indices for each identity dimension, scaled so that higher values indicate a closer match between leaders and the population for that trait.
• Representation Gap: the difference between observed representation and the random-selection benchmark; values closer to zero indicate smaller gaps, more negative values indicate larger shortfalls, and positive values (which do not appear in the current rankings) would indicate overrepresentation of some groups.

For example, a country with a gender representation index of 0.80 has a leadership that is substantially closer to gender parity than a country scoring 0.60, but if both have strongly negative representation gap values they still fall short of what we would expect under random selection from the population.

How Did We Calculate These Country Rankings?

These country rankings are based on the Global Leadership Project (GLP), which compiles detailed information on national political leaders around the world. For each country, we link leaders’ demographic characteristics—such as gender, ethnic affiliation, religious group, and language—to population-level distributions and compute descriptive representation indices for each identity dimension.

We then aggregate these dimension-specific indices into an overall descriptive representation score and calculate the representation gap by comparing observed representation to a random-selection benchmark. Countries are included only if the underlying GLP data reach at least a 75% completeness threshold for the demographic traits we analyze. For cases with limited missingness, we use a principled imputation procedure to fill in missing leader attributes so that countries remain comparable.

The current release covers 156 countries that meet these data-quality criteria. While cross-national patterns are robust, small differences between countries should be interpreted with caution, particularly when scores are close to one another or rely on imputed values. For full methodological details—including the exact construction of each index and the statistical model underlying the imputation—see our American Political Science Review article The Composition of Descriptive Representation.

What Are the Main Global Patterns in Descriptive Representation?

Three broad patterns stand out from the country rankings: representation gaps exist in every country, some countries make much more progress than others, and gender remains the deepest fault line in descriptive representation. The rankings also show that institutional design and social structure—such as legislature size and identity fractionalization—shape how easy it is for leaders to mirror the citizens they represent.

• All countries have lower representation than under random selection. In the current dataset, every country has a negative representation gap score, indicating that political leaders are less diverse than citizens when compared to the random-selection benchmark. Even the best-performing countries, with representation gaps around −0.04, fall short of full demographic parity.
• Some countries come much closer to demographic parity than others. Countries such as Iceland, Norway, Denmark, Finland, and Poland have overall descriptive representation scores around 0.86–0.88 and relatively small representation gaps, indicating comparatively representative leadership. By contrast, countries like Congo, Indonesia, Solomon Islands, Central African Republic, and Sierra Leone sit near the bottom of the rankings, with overall scores closer to 0.50–0.58 and much larger representation gaps.
• Gender is most often the weakest dimension of representation. In a majority of countries, the gender index is the lowest of the four identity dimensions, reflecting the systematic underrepresentation of women in national politics. Even where ethnic, religious, or language representation looks relatively strong, gender representation gaps typically remain large.
• Larger political bodies tend to have smaller gaps than small elites. In our accompanying article, we show that legislatures and leadership bodies with more seats are better able to mirror population diversity than tightly restricted elites, producing smaller representation gaps at the country level. This contrast between large and small institutions helps explain why some countries perform better on descriptive representation despite facing similar underlying social divisions.
• More fractionalized societies face greater challenges. Countries with many distinct ethnic, religious, or language groups must divide a fixed number of leadership positions among more communities, which makes descriptive representation harder to achieve. More homogeneous societies, by contrast, can more easily approach parity, even without ambitious institutional reforms.

How Should Researchers and Policymakers Use These Rankings?

The representation gap index and descriptive representation scores are best used as tools for benchmarking and comparison rather than as simple pass–fail grades. They allow you to see where a country sits relative to others, which identity dimensions are most underrepresented, and how institutional reforms or social changes affect those gaps over time.

• Benchmark gender, ethnic, religious, and language representation by country. Quickly identify where a country stands on each dimension, and whether its representation gaps are unusually large or small compared with peers in the same region or income group.
• Evaluate reforms and quota policies. Track whether gender quotas, minority reserved seats, or changes in electoral rules are associated with smaller representation gaps in legislatures and other leadership bodies.
• Design comparative case studies. Use the rankings to select most-similar or most-different country pairs on descriptive representation, then dig into the institutional and historical reasons behind those contrasts.
• Support teaching, journalism, and public communication. Incorporate the country rankings into classroom exercises, policy briefs, or news stories that explain where descriptive representation is improving or stagnating and why.
• Best for: quick, cross-national benchmarking of political representation gaps. The dataset is intentionally simple and standardized so it can plug into broader research on democratic quality, inequality, and institutional design.

Frequently Asked Questions About the Representation Index